Network structure, social personalities, and the spread of invasive species

Background/Question/Methods Ecological invasions where non-native species spread to new areas, grow to high densities, and have large, negative impacts on ecological communities are a major worldwide problem exacting tremendous economic and ecological costs. A key mystery in invasion ecology is the apparently idiosyncratic nature of the invasion process. Of the many species that are introduced into any given area, only a few establish and of those that establish, only a few spread and grow to pest status. Ecologists have attempted to explain this variability by searching for combinations of ecological conditions and species' traits associated with invasion success; however, this effort has met with only mixed success. Here, we use a mathematical model that integrates insights from our growing understanding of behavioral syndromes (aka animal personalities) with network theory and spatial ecology to derive a new mechanism for explaining variation in animal invasion success. In our model, individual fitness is determined by local patch density, with asocial individuals outperforming social individuals at low density, social individuals performing better at high density. Dispersal between patches is assumed to be negatively related to the individual's expected fitness in the patch. Spread speed, defined as the inverse of the time taken to spread through the network, was evaluated for a range of network structures with either all social, all asocial, or mixed behavioral-type introduction. Results/Conclusions We show that the rapid, epidemic-like spread of occurs most readily when a species includes a mix of social and asocial behavioral types that differ in density-dependent dispersal tendencies, and when the pattern of habitat connectivity is more highly heterogeneous. Within-species variability in behavioral types, a feature not included in previous models, is important for overcoming the fact that different traits are associated with success in different stages of the invasion process. Mixed type invasions spread faster than either monotype, as local growth of social types forces accelerated spreading by asocials at the invasion edge. Additionally, growth to nuisance densities is assured by the presence of social individuals. The results hold implications for the prediction of invasion impacts, as well as the classification of traits associated with invasiveness. If we ignore the possibility of behavioral succession in invasions, we are likely to predict that traits of the climax type facilitate invasion, when in reality they are unlikely to spread without facilitation by an opportunistic edge type.